Pushing the Photon Limit: Nanoantennas Increase Maximal Photon Stream and Total Photon Number

Nanoantennas are well-known for their effective role in fluorescence enhancement, both in excitation and emission. Enhancements of 3–4 orders of magnitude have been reported. Yet in practice, the photon emission is limited by saturation due to the time that a molecule spends in singlet and especially triplet excited states. The maximal photon stream restricts the attainable enhancement. Furthermore, the total number of photons emitted is limited by photobleaching. The limited brightness and observation time are a drawback for applications, especially in biology. Here we challenge this photon limit, showing that nanoantennas can actually increase both saturation intensity and photostability. So far, this limit-shifting role of nanoantennas has hardly been explored. Specifically, we demonstrate that single light-harvesting complexes, under saturating excitation conditions, show over a 50-fold antenna-enhanced photon emission stream, with 10-fold more total photons, up to 108 detected photons, before photobleaching. This work shows yet another facet of the great potential of nanoantennas in the world of single-molecule biology.

P lasmonic nanoantennas are metallic nanoparticles that are resonant at optical frequencies. 1 This leads to localization and concentration of the electromagnetic field into subdiffraction-limited volumes. 2−5 It is well-known that chromophores placed in the localized "hot spot" of such an antenna can show strong emission enhancement. This is usually mostly due to excitation enhancement. However, for low quantum yield emitters, enhancement of the radiative rate can also lead to emission enhancement. The combination of enhanced excitation and emission can lead to 500-fold brighter emission of single light-harvesting complexes 6 and even 1000-fold enhanced fluorescence of dye molecules. 7−9 On the basis of the large enhancements, high quantum yields and short excited-state lifetimes of less than 100 ps 5,6,8 superemitters with 10 10 counts per second could be expected. However, so far, this has not been observed. In practice, the maximal photon emission rate is not limited by the radiative rate to the ground state but by triplet states and photodissociation. The intersystem crossing (ISC) rate and bleach rate limit the photon count rate and number of emitted photons. Strangely enough, while the role of antennas to enhance the excitation and emission rates is widely appreciated, only a few single-molecule studies investigated the antenna enhancement of the typical number of emitted photons N; a 4fold increase was reached by linking a chromophore to gold nanospheres, 10 and a 3-fold increase was found for fluorescent proteins in the presence of gold nanorods. 11 A large increase of the number of emitted photons was reported for single chromophores on gold sphere multimers, 12 but in this work, the characteristic total photon number N was not evaluated. Even more striking, it seems that the role of nanoantennas in enhancement of saturation levels has been largely overlooked.
In theory, a nanoantenna can enhance the emission saturation level of a molecule, as is demonstrated in the following equations. For a molecule excited with low power in the absence of a dark state, the detected photon count rate (PCR, s −1 ) is given by eq 1, with κ the detection efficiency of the microscope, σ the molecular absorption cross section (cm 2 ), I E the excitation intensity (W cm −2 ), hν the photon energy (J), k r the radiative rate (s −1 ), and k tot the total decay rate (s −1 ).
When the excitation power I E is increased, and certainly with strong antenna enhancement, saturation effects start to occur and the equation changes into eq 2, with I Sat as the saturation intensity 13,14 The saturation intensity is at the crossover between the linear and saturated regime, and at I Sat , the PCR is half of the theoretical maximum reached at infinite excitation intensity. In the presence of a dark triplet state, with k ISC the intersystem crossing rate from the singlet excited state to the triplet state (s −1 ), k tot = k r + k nr + k ISC (s −1 ), and k d the decay rate from the triplet state to the ground state (s −1 ), the saturation intensity is given by eq 3.
In the vicinity of a nanoantenna, k tot is larger due to enhancement of k r and k nr ; as a result, saturation will be reached at higher excitation intensities. Substitution of eq 3 in eq 2 gives the PCR at the saturation excitation intensity (PCR Sat ).
Note that the maximum PCR for I E ≫ I Sat is twice PCR Sat . PCR Sat scales linearly with the radiative rate k r and thus with the Purcell factor for a molecule located in the hot spot of an antenna. Purcell factors as high a 600 at a wavelength of 1000 nm have been theoretically predicted for gold nanoantennas, 15 indicating that the saturation count rate can be substantially enhanced.
It has been shown that the emission of single LH2 complexes can be strongly enhanced with nanoantennas, profiting from both excitation and quantum yield enhancement. 6,16−18 Excitation enhancement is extremely useful for selective excitation of complexes present in the antenna hot spot, but it does not increase the maximum photon count rate. In this work, we investigate antenna enhancement of the PCR at saturating excitation intensities. The fluorescence intensity of single LH2 complexes in the absence and presence of gold nanorod antennas was studied as a function of the excitation intensity. To study single LH2s under control conditions, the complexes were diluted to 83 pM in an aqueous poly(vinyl alcohol) (PVA) solution and spin-casted over a glass coverslip ( Figure 1A). The fluorescence as a function of the excitation intensity (linear polarized light, λ = 800 nm) is plotted in Figure 1B. The standard deviation is rather large, most likely due to variations in the orientations and environment of the individual LH2 complexes. Fitting the data shows that saturation occurs at around 92 W cm −2 , with a PCR Sat of 6.4 × 10 3 counts s −1 and thus a maximal PCR for I E ≫ I Sat of 13 × 10 3 counts s −1 . In a second approach, a 4000-fold increased LH2 concentration was used to allow the excitation of a large

The Journal of Physical Chemistry Letters
Letter number of complexes per diffraction-limited spot. In this ensemble experiment, variations between different areas were negligible. A higher saturation intensity was expected because the LH2 complexes were spread over the Gaussian intensity distribution of the excitation spot, while the single complexes were measured in the center of the spot at the highest intensity. This effect is partly compensated by the on/off blinking of single molecules. The off-switching rate is linearly dependent on the excitation intensity, while the on-switching rate is lightindependent. 19 The off states are separated out in the singlemolecule experiments, while they contribute to the ensemble measurement giving rise to an earlier apparent onset of saturation. The ensemble saturation occurred at 135 W cm −2 , in good agreement with the single-molecule measurements.
Next we focus on antenna-enhanced LH2 emission. LH2 complexes were diluted in a PVA solution to a concentration of 6.7 nM and spin-cast over e-beam lithography fabricated arrays of gold nanorod antennas, with dimensions of 50 nm × 60 nm × 160 nm, spaced 1 μm × 1 μm. Light of λ = 800 nm polarized along the long axis of the antenna was used for excitation. This light is resonant with the antenna and leads to enhanced excitation in the antenna hot spots. The LH2 concentration was chosen such that only in a fraction of the antenna was a single LH2 complex located in the hot spot, resulting in bright fluorescence emission 6 ( Figure 1C). The single complexes were identified by the typical blinking behavior of their fluorescence (Figure 2), and their single photon emission behavior was confirmed by photon-antibunching, as we have shown before. 6 The fluorescence intensity as a function of the excitation power for a number of antenna-enhanced LH2 complexes is shown in Figure 1D. The saturation intensity ranges from 4 to 18 W cm −2 . Note that these are far-field intensities; the near-field intensities in the hot spot of the antenna can be 100-fold higher. 6 The PCR Sat ranges from 61 × 10 3 to 340 × 10 3 counts s −1 , meaning that the antenna enhancement of PCR Sat ranges from 10 to 53 times. PCR Sat is linearly dependent on the radiative rate enhancement (eq 4), which is in turn strongly dependent on the orientation and distance of the chromophore dipole with respect to the antenna. On the basis of Finite Difference Time Domain simulations, a maximum radiative rate enhancement of 150-fold was predicted for an optimally orientated dipole (for LH2 emission at λ = 870 nm) located 5 nm from the antenna, while at 25 nm, the enhancement was decreased to 25-fold. 6 Thus, the experimentally observed enhancement of PCR Sat is in the range expected from theory.
All single-molecule experiments are limited by the number of photons emitted before photobleaching (N). Increasing this number with the use of nanoantennas would be extremely valuable. In the simplest situation, photobleaching is a spontaneous decay process from the singlet excited state, occurring with rate k b . In this case, the typical number of emitted photons (N) before bleaching is given by eq 5 and scales linearly with the Purcell factor. 10 However, for a large number of organic chromophores, the bleaching process is more complex and, for instance, involves triplet states. 20−22 Under the relative high irradiance conditions used for single-molecule experiments, molecules in the first electronic excited singlet or triplet states, S 1 and T 1 , may absorb a second photon and reach higher electronic states, S n and T n . These higher states are readily subjectable to photobleaching. 20 The number of emitted photons is no longer independent of the excitation intensity when bleaching occurs through such multiphoton processes. Instead, the highest number is reached at the lowest excitation power. 20,23 For such molecules, N will still increase with the Purcell factor as the molecules spend less

The Journal of Physical Chemistry Letters
Letter time in the excited state when k r is enhanced, but the relation is no longer linear.
We investigated how the nanoantennas enhance the photostability of single LH2 complexes. Examples of fluorescence traces from single LH2 complexes are shown in Figure 2, with both no antenna (LH2) and antenna-enhanced (LH2 + antenna). The traces show the typical single-molecule on/off blinking and finally irreversible photobleaching. For some LH2 complexes, the PCR fluctuates between bright and dimmer states ( Figure 2C,F). Such fluctuations have been seen before and were suggested to arise from different conformational states 24,25 or changes in the radiative rate of the LH2 complex and the formation of a photochemical product with a low probability to trap the excitation. 19 The LH2 complexes photobleach after seconds (Figure 2A,B), minutes ( Figure  2C,D), or even hours ( Figure 2E,F), with cases of up to 10 8 photons detected.
The total number of detected photons for each complex was plotted against the maximum photon count rate ( Figure 3A), and photon number histograms were built ( Figure 3B,C). Fitting of the histogram with an exponential decay function gives the statistical number of photons detected (N Det ), which relates to the bleach rate. Histograms were built for three excitation intensities, one close to the saturation intensity and two below. N Det decreased with increasing excitation powers, both in the absence and in the presence of the antenna ( Figure  3B,C), indicating that photobleaching of the LH2 complexes occurs through a multiphoton process. 20,23 In Figure 4, N Det is plotted against PCR averaged for all of the complexes measured with the same excitation intensity. Higher PCRs come at the cost of a decreased N Det . However, in the presence of the antenna, both the PCR and N Det are approximately 10 times higher compared to the control situation. As such, a single molecule can be measured for the same amount of time with 10-fold higher PCR, allowing processes to be monitored at a 10-fold higher sampling rate.
We did not measure the antenna-enhanced N Det at the same low PCRs (4 × 10 3 to 9 × 10 3 counts s −1 ) as assessed for LH2 without antenna. However, roughly extrapolating the data indicates that under these conditions, N Det would be about 2 orders of magnitude larger. This provides the ideal situation to follow single molecules for a very long time.

The Journal of Physical Chemistry Letters
Letter Single-molecule techniques have found applications in a wide range of life science research fields, including DNA sequencing, 26 super-resolution imaging, 27,28 and photosynthesis. 29−35 However, the maximal number of photons that a single complex can emit per second is restricted by its intrinsic properties. In addition, the total number of photons that a single complex can emit, especially at physiologically relevant temperatures, is limited. This hinders the observation of fast changes and fluctuations in the complex and sets the boundary for the total measurement time. Clearly, a general method that alleviates these restrictions would be valuable. Here we demonstrate an over 50-fold nanoantenna enhancement of the maximum photon count rate from a single light-harvesting complex (LH2). The total number of photons that a single LH2 complex emits was shown to depend on the excitation intensity. The average photon count rate and the total number of detected photons were simultaneously enhanced 10 times by the nanoantenna at approximately 40-times lower laser intensities. These intense enhancements show the great potential of nanoantennas for photosynthesis research in particular and single-molecule biology in general.
Confocal Microscopy. Microscopy was performed using a commercial time-resolved confocal microscope (Micro Time 200, PicoQuant, Germany). The excitation was with linearly polarized pulsed light at λ = 800 nm (Titanium−Sapphire pulsed laser, Coherent-Mira), with a repetition rate of 76 MHz. A high numerical aperture (1.46, 100×, Zeiss) oil immersion objective mounted on an inverted microscope (Olympus) was used for both excitation and collection. The fluorescence light was separated from the excitation light using a dichroic mirror and long-pass filters (λ = 835 nm + 850 nm) and detected by an avalanche photodiode (MPD, Micro Photon Devices).